OpenAI to Watermark ChatGPT Text in EU Under AI Act

6 Oct 2026, 2:06 am · 3h ago · 1 min read · TechCrunch AINew

OpenAI has announced plans to implement text watermarking for ChatGPT and Codex outputs across the European Union to comply with the newly established EU AI Act regulations. The company intends to embed invisible identifiers within generated text to verify synthetic content. However, OpenAI noted that manual editing, paraphrasing, or translating the text could make these digital watermarks significantly harder to detect. The regulatory compliance move marks a major shift in how foundation model providers handle transparency, machine-generated media traceability, and compliance standards within heavily regulated international markets across Europe.